Does Google Look @ the Text Surrounding Links ?

[Sebastian]

Hey,

Say there is a blurb reviewing your site on your other site.

It talks about your site's theme, which is blue widgets and links to your site using your brand widgetsreview.com.

Do you think Google looks at the immediate context of a link ?

So that if enough sites are linking to your generic brand surrounded by copy about blue widgets, it stars associating the two ?

Seb

[Spry]

while i have no hard data to back this up. I believe what you suggest is correct. They do look at the context of the text around your link.

[Shay]

I think Google looks at the context around the link, the anchor, and the entire article for that matter in order to determine relevance and thus association.

[cernus11]

Agreed - Pretty much all my links I built are in context regardless. The way I see it is if they don't take it in to account now they definitely will in the future. In my opinion in context links point towards more naturally looking link portfolio - a lot of out of context links seems very artificial to me.

Like Spry said though - no real hard data but I think they will take it into account now.

[Michael Martinez]

They have published some patents and technical papers that explain how they could look at context for links and why they might do it. They have never fessed up, so far as I can recall, to using that context in their algorithm.

A few years ago I ran some tests where it seemed like some sites were passing context in addition to anchor text; and yet, others were not. And there have been hundreds, perhaps thousands of algorithm changes since then so I wouldn't put much stock in those results.

Interestingly, I am seeing context being passed through 301-redirects to a page on Xenite.Org. It looks like a directory listing description for a now-discontinued service has been picked up by the Sitelinks Google displays. To the best of my knowledge, that text has never appeared in any anchor text for the service and it was certainly never included on any pages on my site.

The page receiving the special attention from Google is a generic "Xenite.Org is being updated" page where I have 301-redirected all old URLs for which I have not published new, relevant content. I've been debating whether to block the sitelink.

If you want to see what I am referring to, click here to view this search result. If you see a sitelink for "Internet Authors Network" (which I shut down completely earlier this year), that descriptive text can be found in many directory listings (and I have no idea of where they came from -- I did not submit the service to all those directories).

I would strongly advise anyone who wants to share an opinion on the possible implications of what I just wrote to KEEP IT TO YOURSELF if you want to experiment with any funky ideas. Blurting out something you think you see is the best way possible to ruin what might be a potentially helpful resource.

Not trying to be rude -- I'm just saying, if you can read between the lines, it would be in your best interest to leave whatever you read there unspoken, unshared, unblogged, unforumed, etc.

It's no skin off my back if you do say something. I'm just suggesting people may want to think about what they want to say before they say it.

[Sebastian]

If I were a search scientist I'd design my algorithm to do this (Sebby Algo Trademark)

* If there is descriptive anchor text: Keep it, because it's already relevant.

* If it's branded anchor text or an absolute URL, look at the link context instead to get relative information.

* Then place all the words in direct vicinity of the brand in a temporary bucket.

* Look at other brand mentions, and do the same.

* Cross relate terms in each bucket to extrapolate theme.

The reasoning is that if people are writing long and descriptive anchor text, Google will take it of course, but if no descriptive anchor text is present, it tries to figure it out.

There's prob some weaknesses in this reasoning. But that's how I'd design my algo.

[Michael Martinez]

I, personally, would prefer to see link anchor text dropped from the ranking algorithms. They never should have included it. Larry Page and Sergey Brin really didn't have a clue about the Web when they designed Google. Amit Singhal rewrite the basic code when they hired him in 2000 but he retained some of the dumbest, most easily spammed functions of the original Google algorithm.

People don't get rich because they are smarter than everyone else -- they get rich in spite of their shortcomings and because they work their butts off. It's unfortunate that some people become so influential on their way to wealth that they cause a lot of problems for other people.

Microsoft did that in the late 1980s and early 1990s. I used to have to argue hard to get business owners to switch from Windows to UNIX because they would look in the phone book and find dozens of local "Windows consultants" but maybe only 1 or 2 UNIX consultants. "What if we have a problem with UNIX?" they would say. "There is so little demand for UNIX consultants there are hardly any out there."

The irony in that complaint just blew me away every time I heard it.

So in the 2000s Google popularized link "building" and spam Websites made for advertising, and people called it a great search engine.

You really have to look at the other side of the coin, in my opinion, before you determine what its true worth should be.

[PTC]

From 2009 - How Important Is Latent Indexing?

[flogger]

Latent indexing is not that effective anymore these days.

[Michael Martinez]

No major search engine has ever implemented latent semantic indexing. It is too resource-inefficient and would not deal with parallel idiomatic pathways.